Identification of Board-Certified Plastic Surgeons Using Artificial Intelligence: An Accuracy Assessment
Rohun Gupta, Atharva M Bhagwat, Mei Hainline, Herluf Lund, Brian A Mailey, Sumesh KaswanAbstract
Background
Artificial intelligence (AI), particularly large language models like ChatGPT and Gemini, is increasingly used in healthcare for diagnosis, education, and research. As interest in cosmetic procedures rise, patients face growing difficulty in selecting qualified providers due to unclear credentials and expanding practitioner types.
Objectives
This study evaluates whether AI can help patients accurately identify board-certified plastic surgeons, addressing a current gap in the literature.
Methods
Four leading AI tools—ChatGPT, Perplexity, Gemini, and Claude—were evaluated by asking each to name board-certified plastic surgeons across all fifty U.S. states. Results were verified through independent internet searches and the American Board of Plastic Surgery to confirm certification status. Statistical analysis using GraphPad Prism found differences between tools, with significance set at α = 0.05.
Results
Across 1,000 results, overall accuracy in identifying plastic surgeons was 94.1%, with Gemini performing best, followed by Perplexity, ChatGPT, and Claude, with statistically significant differences. Most non-plastic-surgeons were otolaryngologists and general surgeons, with no difference between AI models in error type distribution. Gemini scored the highest in regards to board certification accuracy, showing significant differences compared to the other models.
Conclusions
As the field of aesthetic procedures grows and more providers enter the market, patients must be able to make well-informed choices when selecting a practitioner. While artificial intelligence offers a promising tool for identifying board-certified plastic surgeons, its potential for error means patients should independently verify the information.